Performance Prediction of Service-Oriented Applications based on an Enterprise Service Bus

Abstract

An enterprise service bus (ESB) is a standards-based integration platform that combines messaging, web services, data transformation, and intelligent routing in a highly distributed environment. The ESB has been adopted as a key component of SOA infrastructures. For SOA implementations with large number of users, services, or traffic, maintaining the necessary performance levels of applications integrated using an ESB presents a substantial challenge, both to the architects who design the infrastructure as well as to IT professionals who are responsible for administration. In this paper, we develop a performance model for analyzing and predicting the runtime performance of service applications composed on a COTS ESB platform. Our approach utilizes benchmarking techniques to measure primitive performance overheads of service routing activities in the ESB. The performance characteristics of the ESB and services running on the ESB are modeled in a queuing network, which facilitates the performance prediction of service oriented applications. This model is validated by an example ESB based service application modeled from real world loan broking business application.